OPENNESS, INVESTMENT AND ECONOMIC GROWTH IN ASIA Dipendra Sinha (Department of Economics, Ritsumeikan Asia Pacific University, Japan and Macquarie University, Australia, Email [email protected]) and Tapen Sinha (Instituto Tecnologico Autonomo de Mexico, Mexico) I. Introduction The purpose of this paper is to explore how growth of openness and the growth of domestic investment contribute to growth of GDP in Asian countries. A large number of papers have looked at the relationship between export and economic growth but until recently no study has looked at the relationship between growth of openness and economic growth (see Sinha and Sinha (1996)). We define openness as export plus import following Summers and Heston (1991). Why should openness and growth be related in a country? Openness has two parts: export and import. Economists have debated the role of export in economic growth at length. There are three channels of connections between economic growth and exports. First, although industrialization is crucial to economic growth, domestic demand may be low. Export provides an outlet for this excess production and generates income (Colombatto, 1990). Second, in the long run, export helps growth because export leads to greater technical progress and more saving (Krueger, 1978). It also improves credit ratings of a country by generating hard currencies and thus makes getting foreign loans easier. Third, export promotion policies improve total factor productivity (Balassa, 1978). Although most researchers talk about trade policy but in their discussion they focus exclusively on export policy. Does that mean that import does not help a country? Why should import be related to economic growth? Publication of high import statistics in the newspapers stirs up government officials. They feel compelled to defend high import statistics. There is an implicit belief that for a country, a growing rate of import is bad but a growing rate of export is good. Economics literature seems to follow the same line: issues of trade are always implicitly taken to mean issues of export. The only study to our knowledge that explicitly looks at import at all is that of Ram (1990). Ram looks at the relationship between growth rate of import and growth rate of real GDP in many developing countries using an augmented production function approach. He finds a positive relationship for some countries. Import of capital goods and energy can help economic growth for an LDC. However, imports may not always aid economic growth: they need to be used efficiently. We also add two very important internal factors affecting the rate of economic growth: growth of investment and growth of population. Population is used as a proxy for the labor force since reliable time series data are not available for a reasonable period of time for all the countries under consideration. Our model in section II derived from a generalized version of the Solow-Swan model makes growth of GDP a function of growth of investment and the growth of population as well. We use data for 19 Asian countries. World Bank (1987) categorizes a number of countries according to their trade orientation for the period 1973-85. Our sample includes countries in all these categories. Hong Kong, Singapore and South Korea were strongly outward oriented according to the World Bank classification. Israel, Malaysia and Thailand were moderately outward oriented. Indonesia, Pakistan, Philippines and Sri Lanka were moderately inward oriented. Bangladesh and India were strongly inward oriented. Therefore, our data span the entire range of countries classified by the World Bank. As pointed out earlier, previous literature look solely at the relationship between export and economic growth. Early efforts investigating the relationship between export and economic growth include Emery (1967), Michaely (1977), Balassa (1978), Krueger (1978) and Feder Volume 49, No.4 91 (1982). Numerous other studies also appeared on the subject. Some of these are multi-country studies while others concentrate on a single country. Recent papers include Ahmed and Harnhirun (1995), Dollar (1992), Harrison (1995), Frankel, Romer and Cyrus (1995), Krueger (1990), Sengupta (1994) and van den Berg and Schmidt (1994). Edwards (1993) provides an excellent review of the many previous studies. The plan of the paper is as follows. In part II, we develop a model relating the rate of growth of GDP, the rate of growth of investment, the rate of growth of openness and the rate of growth of population. In part III, we look at the empirical results. Part IV gives a summary of the paper and draws some conclusions. II. The Model Define the following variables: Y(t) = real GDP adjusted for terms of trade as a (continuous) function of time; K(t) = capital stock at time t; N(t) = labor in person-hours in efficiency units at time t; L(t) = population at time t; T(t) = labor productivity or technical change at time t; R(t) = value of trade (real export plus real import) at time t; We define a continuous time model of the economy as follows: There is a continuous time technology given by the following function: Y(t) = F(K(t), N(t)) (1) and we assume F is homogenous of order 1 in K and N. This formulation is the standard Ramsay-Solow-Swan model but in continuous time. We will use the following form of (1) by using the fact that F is homogeneous of degree 1: Y(t)/N(t) = f(K(t)/N(t)) (1) Capital stock evolves over time as a proportion of real GDP but after correcting for depreciation at a constant rate d. This is the simplest formulation of Solow (1956). dK(t)/dt = sY(t) - dK(t) (2) We assume that the population grows at a constant rate n over time: dL(t)/dt = nL(t) (3) Labor productivity/technology is assumed to change with improvements in per capita capital stock K/L. This is an extension of Arrow’s(1962) learning by doing model where a is the learning coefficient. It also captures the spirit of endogenous growth model of Romer (1986). We will assume that a is influenced by import and export. There are two separate sets of reasons why import and export (or trade) induces a higher value of a. First, more sophisticated imported technology stimulate output growth (see Bardhan and Lewis (1970)). Feder (1982) shows how export can produce a higher level of productivity. Second, a purely domestic source stemming from local technological improvement can be proxied by export and a foreign source of knowledge related to innovations generated in other countries can be proxied by imports (see Edwards (1992) for more details on the rationale of using export and import as proxies in this context). Let R(t) denote trade (that is, real export plus real import). In our model the parameter a encapsulates all of these effects of import and export. dT(t)/dt = a(dR(t)/dt)K(t)/L(t) + lT(t) (4) Equation (4) summarizes the rate of technological change. In the standard neoclassical growth model, the technological change is exogenous. This exogeneity means equation (4) simplifies to: dT(t)/dt = lT(t) (4¢) This is equivalent to assuming the value of to be zero. In this sense, (4) is a generalization of Solow-Swan neoclassical model (Solow (1956), Swan (1956)). In addition, we have the following identities: N(t) = T(t)L(t) (5) Equation (5) relates population to effective labor. k(t) = K(t)/N(t) (6) Equation (6) defines capital intensity. From (2) using (1), we get dK(t)/dt = sNf(k(t)) - dK(t) which simplifies to 92 THE INDIAN ECONOMIC JOURNAL [dK(t)/dt]/K(t) = sf(k(t))/k(t) - d (7) Equation (7) summarizes the rate of change of capital stock to per capita capital stock . Differentiating equation (5), we get, dN(t)/dt = L(t)dT(t)/dt + T(t)dL(t)/dt; which after rearranging gives us [dN(t)/dt]/N(t) = [L(t)/N(t)]dT(t)/dt + [T(t)/N(t)]dL(t)/dt (8) By substituting (3) and (4) in the right hand side of (8), we get [dN(t)/dt]/N(t) = [a(dR(t)/dt)K(t)/L(t) + l T(t)] + nL(t)/L(t) or, [dN(t)/dt]/N(t) = a(dR(t)/dt)k(t) + l + n (9) Differentiating equation (6) with respect to t we get dk(t)/dt = [dK(t)/dt]/N(t) - K(t)[dN(t)/dt]/(N(t))2 This gives the following expression for [dk(t)/dt]/k(t) = [dK(t)/dt]/K(t) - [dN(t)/dt]/N(t) (10) By substituting (7) and (9) in (10), we get [dk(t)/dt]/k(t) = sf(k(t))/k(t) - a(dR(t)/dt)k(t) - l - n - d; On the other hand, using (1¢) we get Y(t) = N(t)f(k(t) (11) and using (5) in (11) we get Y(t) = T(t)L(t)f(k(t) (12) Totally differentiating (11), we get dY(t)/dt = [dT(t)/dt]L(t)f(k(t)) +T(t)f(k(t))[dL(t)/dt] + f’(k(t))dk(t)/dt The right hand side simplifies to yield [dY(t)/dt]/Y(t) = a(dR(t)/dt)K(t)f(k(t))/Y(t) + lT(t)L(t)f(k(t))/Y(t) + T(t)L(t)n/N(t) + f’(k(t))[dk(t)/dt]/Y(t) or, [dY(t)/dt]/Y(t) = a(dR(t)/dt)k(t) + l + n + f’(k(t))[dk(t)/dt]/Y(t) (13) If we assume Inada (1963) condition to hold for the per capita production function f in equation (1’), we can show that a steady state solution k* to (13) exists. Non-steady state properties of (13) have been explored in detail by Villanueva (1994). Note that in a steady state dk(t)/dt = 0. Therefore, (13) can be rewritten as [dY(t)/dt]/Y(t) = a(dR(t)/dt)k*(t) + l + n (14) A general functional formulation of (14) would be [dY(t)/dt]/Y(t) = g(a(dR(t)/dt), n, k) (15) Equation (15) says that growth rate of GDP is a function of the (a) the learning function, (b) the rate of growth of investment rate and (c) the growth rate of population. Since the learning rate is a function of the growth rate of trade (export plus import in real terms), a linearized version of (15) can be written as: gyt = a0 + a1gxmt + a2git + a3nt + et (16) where gyt is the growth rate of real GDP, xmt is the growth rate of openness, git is the growth rate of real investment, and nt is the population growth rate. This version of our model will be used in our empirical analysis. III. Data and Empirical Results All data come from the Penn World Table (version 5.6). Annual data are used as follows: Bangladesh (1959-92), China (1960-92), Hong Kong (1960-92), India (1950-92), Indonesia (1960-92), Iran (1955-92), Iraq (1953-87), Israel (1953-92), Japan (1950-92), Jordan (1954-90), Malaysia (1955-92), Myanmar (1950-89), Pakistan (1950-92), Philippines (1950-92), Singapore (1960-92), Sri Lanka (1950-92), South Korea (1953-91), Taiwan (1951-92) and Thailand (1950- 92). The Penn World Table data, which were developed by the International Comparison Project, have data, which are much more useful for multi-country studies because these data are more comparable than data from any other sources.

It is important to test for stationarity of the variables involved before attempting empirical estimates. We use the Phillips-Perron (1988) test. The test is well suited for analysing time series whose differences may follow mixed ARMA (p,q) processes of unknown order in that it Volume 49, No.4 93 the test statistic incorporates a nonparametric allowance for serial correlation in testing the regression. Consider the following equation: ~ ~ ~

yt = c 0 + c 1 yt-1 + c 2 (t - T/2) + nt (17) where T is the number of observations and nt is the error term. The null hypothesis of a unit root ~

is: c 1 =1. The we Table 1. Phillips-Perron Unit Root Tests can drop the trend Country gyt gxmt git nt W. T. Trend W. T. Trend W. T. Trend W. T. Trend term to test the stationarity of a Bangladesh -11.21 -12.56 -7.19 -7.14 -6.66 -6.63 -10.41 -11.40 (-2.62) (-3.21) (-2.62) (-3.21) (-2.62) (-3.21) (-2.62) (-3.21) variable without the China -3.72 -3.70 -5.73 -6.95 -6.18 -4.12 -4.16 -5.88 trend. (-2.62) (-3.21) (-2.62) (-3.21) (-2.62) (-3.21) (-2.62) (-3.21) Hong Kong -4.04 -4.55 -6.55 -6.81 -3.49 -3.51 -4.67 -5.39 The results of the (-2.62) (-3.21) (-2.62) (-3.21) (-2.62) (-3.21) (-2.62) (-3.21) India -6.17 -6.64 -6.55 -7.16 -9.05 -9.53 -3.19 -3.32 PP tests are in Table (-2.60) (-3.19) (-2.60) (-3.19) (-2.60) (-3.19) (-2.60) (-3.19) 1. The results Indonesia -3.38 -3.44 -8.25 -9.51 -2.91 -5.81 -2.77 -3.34 (-2.62) (3.21) (-2.62) (3.21) (-2.62) (3.21) (-2.62) (3.21) indicate that all Iran -4.68 -4.70 -3.23 -3.18 -5.77 -5.76 -5.22 -5.88 (-2.61) (-3.20) (-2.61) (-3.20) (-2.61) (-3.20) (-2.61) (-3.20) variables are Iraq -4.80 -4.94 -4.95 -4.98 -4.49 -4.41 -3.14 -3.69 stationary for 15 (-2.61) (-3.21) (-2.61) (-3.21) (-2.61) (-3.21) (-2.61) (-3.21) Israel -3.95 -4.12 -5.61 -6.23 -4.94 -4.86 -3.59 -3.76 countries. However, (-2.61) (-3.20) (-2.61) (-3.20) (-2.61) (-3.20) (-2.61) (-3.20) for four countries, Japan -3.45 -4.31 -8.41 -9.88 -6.70 -6.96 -3.87 -5.45 (-2.60) (-3.19) (-2.60) (-3.19) (-2.60) (-3.19) (-2.60) (-3.19) namely, Malaysia, Sri Jordan -6.09 -6.53 -6.91 -6.85 -5.29 -6.00 -5.84 -5.82 (-2.61) (-3.20) (-2.61) (-3.20) (-2.61) (-3.20) (-2.61) (-3.20) Lanka, Taiwan and Malaysia -4.78 -4.71 -4.83 -5.00 -6.37 -6.29 -1.57 -2.12 Thailand, at least one (-2.94) (-3.20) (-2.94) (-3.20) (-2.94) (-3.20) (-2.94) (-3.20) Myanmar -6.44 -6.31 -4.92 -4.78 -7.55 -7.59 -5.10 -5.52 of the two test (-2.61) (-3.20) (-2.61) (-3.20) (-2.61) (-3.20) (-2.61) (-3.20) statistics for nt (that Pakistan -6.04 -6.11 -6.92 -7.06 -7.26 -7.47 -5.75 -6.21 (-2.60) (-3.19) (-2.60) (-3.19) (-2.60) (-3.19) (-2.60) (-3.19) is, without trend and Philippines -3.16 -3.82 -6.24 -6.19 -4.85 -4.93 -5.07 -5.79 with trend) shows the (-2.60) (-3.19) (-2.60) (-3.19) (-2.60) (-3.19) (-2.60) (-3.19) Singapore -3.56 -3.48 -4.28 -4.18 -4.37 -4.41 -5.05 -4.96 presence of a unit (-2.62) (-3.21) (-2.62) (-3.21) (-2.62) (-3.21) (-2.62) (-3.21) South -4.64 -4.80 -6.34 -8.05 -1.99 -9.89 -7.31 -8.20 root. Thus, we Korea (-2.61) (-3.20) (-2.61) (-3.20) (-2.61) (-3.20) (-2.61) (-3.20) exclude these Sri Lanka -8.20 -8.26 -5.87 -5.93 -7.51 -7.45 -1.96 -5.49 (-2.60) (-3.19) (-2.60) (-3.19) (-2.60) (-3.19) (-2.60) (-3.19) countries from our Taiwan -5.41 -5.33 -6.00 -6.06 -6.25 -6.23 -2.00 -6.10 regression analyses. (-2.61) (-3.20) (-2.61) (-3.20) (-2.61) (-3.20) (-2.61) (-3.20) Thailand -5.41 -5.33 -6.00 -6.06 -6.25 -6.23 -2.00 -6.10 The results of (-2.60) (-3.19) (-2.60) (-3.19) (-2.60) (-3.19) (-2.60) (-3.19) Note: W. T. and Trend stand for test statistics for a variable without trend and with trend respectively. Critical regressions are given values (from Mackinnon (1991)) at 10% level are in parentheses. in table 2. Countries that have followed vigorous trade and (domestic and foreign) investments (China, Hong Kong, Indonesia, Japan etc.) show a very high adjusted R2. These numbers are surprisingly low for Singapore and South Korea. Unfortunately, we could not carry out similar analysis for Taiwan as the population growth rate for Taiwan exhibit unit root. For a number of countries (Bangladesh, India, Iran, Iraq, Japan, Pakistan and Singapore), population growth rate has a negative coefficient contrary to what we predict from our theory. However, none of these coefficients are significantly different from zero. The only country with a significantly negative coefficient for population growth rate is South Korea. Growth rate of investment and of openness have been insignificant among the countries which have pursued either inward looking policies or policies that have punished local investment through stifling regulation. For example, South Asian countries Bangladesh and India show very small impact of either openness or investment. Pakistan shows a negative impact of investment growth. However, openness or trade (mainly with other Muslim countries in the region) has had positive impact on the growth in Pakistan. But due to high taxation, government intervention and regulation of domestic companies, high investment did not result in higher growth. Most of the investment went to unproductive activities. This is 94 THE INDIAN ECONOMIC JOURNAL

Table 2. Regression Results particularly true for the late 1960s and Error Adj. Country gxm gi n a 2 Proce 1970s when the public sector t t t 0 R ss investment expanded rapidly. Many 0.130 0.004 -0.224 0.042 Bangladesh 0.03 OLS (1.972) (0.072) (-0.11) (0.743) industries were nationalized during this 0.088 0.267 0.514 0.014 China 0.71 OLS period and were very badly managed (2.186) (5.703) (0.444) (0.675) 0.152 0.277 0.316 0.042 Hong Kong 0.66 OLS (Sinha (1997)). These results stand in (2.098) (5.750) (0.912) (3.882) 0.050 0.072 -0.501 0.045 sharp contrast with the ASEAN India 0.01 OLS (0.732) (1.012) (-0.21) (0.851) economies. For Indonesia, Myanmar, -0.006 0.150 5.136 -0.055 Indonesia 0.65 AR(1) (-0.62) (5.701) (1.712) (-0.90) the Philippines and Singapore, the 0.264 0.022 -0.873 0.058 Iran 0.39 AR(1) positive impact of openness and (4.660) (0.388) (-0.80) (1.442) 0.482 0.013 -5.523 0.201 investment are robust. For Pacific Iraq 0.54 OLS (5.086) (0.203) (-0.62) (0.701) 0.084 0.164 1.112 0.017 Rim countries like China, Hong Kong Israel 0.60 OLS (2.362) (4.829) (2.648) (1.298) and South Korea, the result of higher -0.099 0.195 -0.442 0.055 Japan 0.72 AR(1) (-3.93) (7.089) (-0.51) (3.816) investment and higher trade have led -0.011 0.226 0.283 0.040 Jordan 0.21 OLS to higher economic growth. (-0.13) (3.101) (1.094) (2.156) 0.308 -0.010 1.821 0.006 Myanmar 0.42 OLS (5.383) (-0.40) (0.709) (0.114) IV. Summary and Conclusions 0.146 -0.190 -1.325 0.087 Pakistan 0.36 AR(1) The main purpose of the paper has (4.084) (-4.00) (-2.29) (4.197) 0.047 0.147 0.335 0.024 Philippines 0.55 AR(1) been to study the effects of growth of (1.514) (4.031) (0.400) (1.036) 0.195 openness and investment on the 0.173 -0.224 0.049 Singapore (3.068) 0.42 OLS (3.337) (-0.58) (3.922) growth of GDP. First, we develop a 0.058 0.084 -2.065 0.143 model relating the growth of GDP to South Korea 0.42 AR(1) (1.417) (3.858) (-2.04) (4.595) the growth rates of openness (export Note: T values are in parentheses. plus import), domestic investment and population. Most previous studies relate economic growth to export. But this is likely to create a missing variable bias in such studies because investment is clearly an important factor affecting economic growth. Next, we use Phillips-Perron unit tests on the data for 19 countries to find out whether the data are stationary or not. For 15 countries, all the variables are found to be stationary. However, for four countries, namely, Malaysia, Sri Lanka, Taiwan and Thailand, the growth of population is found to be non-stationary either for trended or for non-trended or for both cases. We exclude these countries from our regression analysis. Our theory predicts that the growth rate of GDP should be positively related to the growth rates of openness, domestic investment and population. For many countries, the results support the theory. For China, Hong Kong, Iran, Iraq, Israel, Myanmar, Pakistan and Singapore, the coefficient of the growth of openness is positive and significantly different from zero. For China, Hong Kong, Indonesia, Israel, Japan, Jordan, Philippines, Singapore and South Korea, the coefficient of the growth of domestic investment is positive and significantly different from zero. In some cases, the coefficient of the growth of population is negative but in all such cases, it is not significantly different from zero. Thus, generally speaking, we find support for the proposition that the growth rate of GDP is positively related to the growth rates of openness and domestic investment. However, the relationship between the growth rate of GDP and the growth rate of population is not that clear cut. REFERENCES 1. Ahmed, Jaleel and Harnihuran, Somchai (1995) “Unit Roots and Cointegration in Estimating Causality between Exports and Economic Growth: Empirical Evidence from the ASEAN Countries,” Economics Letters, 49, 329- 34. 2. Arrow, Kenneth (1962) “The Economic Implications of Learning by Doing,” Review of Economic Studies, 29, 155-73. 3. Balassa, Bela (1978) “Exports and Economic Growth: Further Evidence,” 4. Journal of Development Economics, 5, 181-89. 5. Bardhan, Pranab, and Lewis, Sydney (1970), “Models of Growth with Imported Inputs,” Economica, 37, 373-85. 6. Colombatto, Enrico (1990) “An Analysis of Exports and Growth in LDCs,” Kyklos, 43, 579-98. Volume 49, No.4 95 7. Dollar, David (1992) “Outward-oriented Developing Economies Really Do Grow More Rapidly: Evidence from 95 LDCs, 1976-85,” Economic Development and Cultural Change, 40, 523-44. 8. Edwards, Sebastian (1992) “Trade Orientation, Distortions and Growth in Developing Countries, Journal of Development Economics, 39, 31-57. 9. Edwards, Sebastian (1993) “Openness, Trade, Liberalization, and Growth in Developing Countries,” Journal of Economic Literature, 31, 1358-93. 10. Emery, Robert F. (1967) “The Relation of Exports and Economic Growth,” Kyklos, 20, 470-86. 11. Feder, Gershon (1982) “On Exports and Economic Growth,” Journal of Development Economics, 12, 59-73. 12. Frankel, Jeffrey, Romer, David and Cyrus, Teresa (1995) “Trade and Growth in East Asian Countries: Cause and Effect?” Center for International and Development Economics Research Paper No. C95-050, University of California, Berkeley, CA. 13. Harrison, Ann (1995) “Openness and Growth: A Time Series, Cross Country Analysis for Developing Countries,” National Bureau of Economic Research Working Paper No. 5221, August. 14. Inada, Ken-Ichi (1963) “On a Two Sector Model of Economic Growth: Comments and Generalization,” Review of Economic Studies, 30, 119-27. 15. Krueger, Anne O. (1978) Foreign Trade Regimes and Economic Development: Liberalization Attempts and Consequences (Cambridge, MA: Ballinger for NBER). 16. Krueger, Anne O. (1990) “Asian Trade and Growth Lessons,” American Economic Review, 80, 108-12. 17. Mackinnon, James G. (1991) "Critical Values for Cointegration Tests", in Engle, R. F. and C W. J. Granger (eds) Long-Run Economic Relationships, Oxford: Oxford University Press, Chapter 13. 18. Michaely, Michael (1977) “Exports and Growth: An Empirical Investigation,” Journal of Development Economics, 4, 49-54. 19. Phillips, Peter C. B. and Pierre Perron (1988), "Time Series Regression with a 20. Unit Root", Biometrika, 75, 335-46. 21. Ram, Rati (1990) “Imports and Economic Growth: A Cross Country Study,” Economia Internazionale, 43, 45- 66. 22. Romer, Paul, (1986) “Increasing Returns and Long Run Growth,” Journal of Political Economy, 94, 1986, 1002-37. 23. Sengupta, Jati K. (1994) “Empirical Tests of New Growth Theory: Openness and Growth,” Journal of Economic Integration, 9, 393-415. 24. Sinha, Dipendra (1997), “Tests of Wagner’s Law for Pakistan: A Cointegration Approach,” Indian Journal of Quantitative Economics, 12, 1-18. 25. Sinha, Tapen and Dipendra Sinha (1996), “An Empirical Investigation into the Relationship between Openness and Economic Growth: Evidence from Asia,” Rivista Internazionale di Scienze Economiche e Commerciali, 43, 359-70. 26. Summers, Robert and Heston, Alan (1991) “The Penn World Table (Mark 5): An Expanded Set of International Comparisons, 1950-88,” Quarterly Journal of Economics, 106, 327-68. 27. Solow, Robert (1956), “A Contribution to the Theory of Economic Growth,” Quarterly Journal of Economics, 70, 65-94. 28. Swan, Trevor (1956), “Economic Growth and Capital Accumulation,” Economic Record, 32, 334-62. 29. van den Berg, Hendrik and Schmidt, James R. (1994) “Foreign Trade and Economic Growth: Time Series Evidence from Latin America,” Journal of International Trade and Economic Development, 3, 249-68. 30. Villanueva, Delano (1994) “Openness, Human Development and Fiscal Policies,” IMF Staff Papers, 41, 1-29. 31. World Bank (1987) World Development Report, 1987 (New York: Oxford University Press).